OpenAI launches ChatGPT Images for AI photo editing
OpenAI is updating ChatGPT Images with a model that edits photos more precisely, preserves details better, and generates images up to four times faster. The feature is coming to all ChatGPT users and the API as GPT Image 1.5, with input and output images costing 20% less.
OpenAI has launched a new version of ChatGPT Images, its tool for creating and editing images with artificial intelligence. The model promises to follow instructions more closely, preserve more photo details, and generate results up to four times faster.
The update is coming to ChatGPT for all users and to the API, where it is available under the name GPT Image 1.5. You do not need to choose a different model: the feature is integrated directly into ChatGPT and is rolling out globally.
More precise edits
The main change appears when you upload an image and ask to modify it. The model tries to change only what you specify while preserving elements such as:
- Lighting and composition.
- People’s appearance.
- Logos and other important visual elements.
- Continuity across multiple edits to the same image.
In practice, you can ask it to change an item of clothing, try a different hairstyle, remove an object, or combine two images without the rest of the scene changing unexpectedly. It also supports more creative changes, such as applying a visual style or turning a photo into an illustration without completely losing its original identity.
The tool can add, remove, combine, blend, or reposition elements. That capability is especially useful when you need to prepare a specific image and do not want to start from scratch each time.
A dedicated space inside ChatGPT
OpenAI has also created a dedicated Images section inside ChatGPT. It is available from the sidebar in the mobile app and at chatgpt.com.
It includes dozens of filters, styles, and ready-made ideas that are updated over time. You can try them without writing a detailed prompt, which is useful if you know what kind of result you want but not how to describe it.
The tool also improves text generation inside images. It can now handle smaller, denser text, as well as compositions with many small faces and more natural-looking results. Even so, OpenAI acknowledges that the system still makes mistakes and has room to improve.
What changes for businesses and developers
In the API, GPT Image 1.5 includes the same improvements to editing and detail preservation. OpenAI says it maintains logos and brand elements more consistently, which can help create campaigns, marketing materials, or product catalogs from a single image.
The price is changing too: input and output images are 20% cheaper than with GPT Image 1. That lets you generate more variations with the same budget, such as different backgrounds, angles, or versions of a product.
For you, the most noticeable difference will be that editing an image should require fewer attempts. The feature does not guarantee perfect results yet, but it points to more practical uses: correcting a photo, preparing a social media asset, or exploring a visual idea without mastering professional tools.
The previous version of ChatGPT Images will remain available as a custom GPT. The next thing to watch is whether OpenAI can maintain this level of precision in increasingly complex edits, especially when an image contains text, people, and many elements that need to remain unchanged.